Gastrointestinal hormonal dysfunction in gastroparesis and functional dyspepsia

Authors

  • J. Khoo,

    1. Centre of Clinical Research Excellence in Nutritional Physiology, School of Medicine, University of Adelaide, Royal Adelaide Hospital, Adelaide, SA, Australia
    Search for more papers by this author
  • C. K. Rayner,

    1. Centre of Clinical Research Excellence in Nutritional Physiology, School of Medicine, University of Adelaide, Royal Adelaide Hospital, Adelaide, SA, Australia
    Search for more papers by this author
  • C. Feinle-Bisset,

    1. Centre of Clinical Research Excellence in Nutritional Physiology, School of Medicine, University of Adelaide, Royal Adelaide Hospital, Adelaide, SA, Australia
    Search for more papers by this author
  • K. L. Jones,

    1. Centre of Clinical Research Excellence in Nutritional Physiology, School of Medicine, University of Adelaide, Royal Adelaide Hospital, Adelaide, SA, Australia
    Search for more papers by this author
  • M. Horowitz

    1. Centre of Clinical Research Excellence in Nutritional Physiology, School of Medicine, University of Adelaide, Royal Adelaide Hospital, Adelaide, SA, Australia
    Search for more papers by this author

Address for Correspondence
Professor Michael Horowitz, Discipline of Medicine, Level 6 Eleanor Harrald Building, Royal Adelaide Hospital, The University of Adelaide, SA 5005, Australia.
Tel: +61 (0)8 8222 5501; fax: +61 (0)8 8223 3870;
e-mail: michael.horowitz@adelaide.edu.au

Abstract

Background  Numerous hormones secreted by the gut, during both the fasted state and in response to a meal, influence gastrointestinal motor and/or sensory function, and appear to contribute to the pathogenesis of delayed gastric emptying associated with gastroparesis, functional dyspepsia (FD) and feed intolerance in critical illness. Gut hormones are, accordingly, potential targets for the management of these patients.

Purpose  This article will discuss the hypersensitivity to enteral fat and endogenous (nutrient-stimulated) and exogenous cholecystokinin (CCK) in patients with FD, and the elevation in both fasting and postprandial CCK levels evident in this group. It will review the use of pharmacological agonists of motilin and ghrelin, which accelerate gastric emptying, in the management of gastroparesis and FD. The frequent finding of markedly delayed gastric emptying in the critically ill will be examined; this is associated with elevated plasma CCK and peptide YY in both the fasted and postprandial states, which may account for the increase in small intestinal nutrient inhibitory feedback on gastric motility in this group. The concepts that the rate of gastric emptying is a major determinant of postprandial glycemic excursions in diabetes, and that modulation of gastric emptying may improve glycemic control, will be addressed; in type 1 and insulin-treated type 2 diabetic patients, co-ordination of insulin administration with nutrient delivery and absorption should be optimized, while type 2 patients who are not on insulin are likely to respond to dietary and/or pharmacological interventions which slow gastric emptying.

Introduction

Gastroparesis, characterized by abnormally delayed gastric emptying associated with upper gastrointestinal symptoms and/or impaired blood glucose control (in insulin-treated diabetes mellitus) in the absence of mechanical obstruction, represents a major cause of morbidity and health care costs.1 Normal gastric emptying is dependent on the integration of motor activity in the proximal and distal stomach, and the proximal small intestine. The contractile activities of the antrum and pylorus are controlled by electrical slow waves generated by the interstitial cells of Cajal (ICC), located in the greater curvature of the stomach. In health, the rate of delivery of nutrients from the stomach into the small intestine is regulated closely at an overall rate of 1–4 kcal min−1,2,3 predominantly as a result of inhibitory feedback arising from the interaction of nutrients with receptors in the lumen of the small intestine. The latter is mediated by both neural and hormonal mechanisms. Numerous hormones secreted by the gut, particularly in response to ingested nutrients, have the capacity to influence gastrointestinal motor and sensory function (Table 1). Hormones released during fasting, which are known to have gastrointestinal activity, include motilin, somatostatin, xenin, orexin A and B, and ghrelin.4 In the postprandial state, the list is even longer and includes gastrin, cholecystokinin (CCK), leptin, enterostatin, peptide YY (PYY), apolipoprotein A-IV, glucagon-like peptide-1 (GLP-1), glucagon-like peptide-2 (GLP-2), glucose-dependent insulinotropic polypeptide (GIP), pancreatic polypeptide, oxyntomodulin and amylin.4

Table 1.   Effects of gut hormones on gastric emptying, appetite and glycemic control
Gut hormoneSite of secretionStimulus for releaseGastric emptyingAppetiteGlycemic control
MotilinStomach, duodenumFasting, associated with phase III of MMCAcceleratesNot knownNot known
GhrelinStomachFastingAcceleratesNot known
Cholecystokinin (CCK)Duodenum, jejunumFat, proteinSlowsNot known
Glucagon-like peptide-1 (GLP-1)Distal small intestine, colonGlucose, fatty acids, amino acidsSlows↑ insulin secretion in response to oral glucose load (‘incretin’ effect)
Glucose-dependent insulinotropic peptide (GIP)Proximal small intestineGlucose, fatty acids, amino acidsNo effect↑ insulin secretion in response to oral glucose load (‘incretin’ effect)
Peptide YY (PYY)Distal small intestineFatty acids, amino acids, glucoseSlowsNot known

Gastroparesis may be either transient or chronic, and is traditionally assumed to result from abnormalities in the gastroduodenal smooth muscle and/or its innervation, although, in most cases, the mechanisms underlying the structural, functional and electrical rhythm abnormalities that are evident are poorly defined. It is also recognized that endocrine or metabolic derangements may occasionally cause gastroparesis. Hyperglycemia and hypokalemia have been associated with delayed gastric emptying and, accordingly, are potentially reversible systemic causes of gastroparesis. Hypopituitarism,5 Addison’s disease (hypoadrenalism),6 and hypothyroidism7 have all been associated with delayed gastric emptying and upper gastrointestinal symptoms. The terminology ‘gastroparesis’ has sometimes been reserved for patients who have longstanding symptoms, and markedly delayed gastric emptying. Diabetes mellitus appears to be the most common cause of chronic gastroparesis, accounting for perhaps one-third of cases.1

This article critically reviews the evidence for, and against, the concept that gastroparesis and functional dyspepsia (FD) can be regarded as ‘hormonal disorders’. Particular attention is given to the role of gut hormones in symptom induction in FD and in gastroparesis in the critically ill, and the impact of gastric emptying on glycemic control in diabetes. Cholecystokinin, PYY and GLP-1 – hormones released from the small intestine in response to the presence of nutrients – and ghrelin, which is released predominantly from the stomach in the fasted state and suppressed by meal ingestion, appear to be of particular importance in the regulation of upper gastrointestinal motility. There is persuasive evidence that these hormones also play a role in the regulation of appetite: acute, exogenous administration of CCK, GLP-1 and PYY decrease, while ghrelin increases, energy intake,8 and it is likely that hunger, satiation and nausea are points on the same physiological spectrum, so that high doses of ‘satiating’ gut hormones have the potential to induce ‘gastroparesis’.

The role of gut hormones in functional dyspepsia

Functional dyspepsia is a heterogeneous disorder of poorly defined etiology: a subgroup of patients experience ‘dysmotility-like’ symptoms of nausea, bloating and early satiation, which are associated with abnormalities in energy intake and gut motor function.9 The latter include initially accelerated (in some patients) and overall delayed (in others) gastric emptying, impaired proximal gastric relaxation, increased perception of gastric distension, and disordered antro-duodenal motility. Patients with FD also exhibit increased sensitivity to mechanical distension of the stomach10 and chemical stimulation of the small intestine,11 and recent evidence indicates that, as a group, they eat fewer meals and consume less energy when compared to healthy individuals.12 Perhaps ∼30% of patients with FD have abnormally slow gastric emptying, although the magnitude of the delay is often modest.9 Whether to label such patients as having ‘idiopathic gastroparesis’, rather than diagnosing them with FD, is an area of controversy. In this review, we have not attempted to discriminate potential hormonal mechanisms relevant to idiopathic gastroparesis from those pertaining to FD, at least in part because there is no specific information relating to the former. Gut hormones have been proposed as mediators for the processes that lead to both the development, and exacerbation, of symptoms in FD. This hypothesis has been strengthened by evidence that the induction of symptoms in FD is ‘nutrient-specific’, so that FD patients appear to be more sensitive to fat than carbohydrate.13

The few studies which have investigated the effects of gut hormones on symptom induction in FD, particularly that associated with fat ingestion, suggest that CCK, secreted from the small intestine in response to the presence of products of fat and protein digestion,8 plays a role in the pathogenesis of FD. Studies demonstrating that acute inhibition of fat digestion by the lipase inhibitor, orlistat,14 and blockade of intestinal chemoreceptors by topical anesthesia with benzocaine,15 reduce dyspeptic symptoms induced by duodenal lipid infusion with, or without, concomitant gastric distension, establish that nutrients (and fats in particular) cause symptoms via mucosal mechanisms in FD, but they did not assess whether symptoms were hormonally and/or neurally mediated. There is evidence, albeit limited, that intravenous infusion of the CCK analogue, CCK-8, induces symptoms in a much greater proportion of patients with FD than healthy volunteers,16 suggesting that the ‘sensitivity’ to CCK may be increased. Moreover, fat-induced dyspeptic symptoms have been reported to be attenuated, both acutely and chronically, by dexloxiglumide.16 Fasting and postprandial CCK concentrations are greater in FD and related to the severity of nausea.13 Limitations inherent in these studies include the relatively small size of the cohorts and the heterogeneous nature of FD. A larger study relating to the safety and efficacy of dexloxiglumide in FD (http://clinicaltrials.gov/ct2/show/results/NCT00303264) was completed in August 2007 and, to the authors’ knowledge, has not been published. It is also important to recognize that the potentially beneficial direct effects on sensation of blocking CCK receptors may be opposed by the associated acceleration of gastric emptying. The induction of dyspeptic symptoms by CCK is likely to be mediated, at least in part, by muscarinic mechanisms, as administration of atropine prevents nausea and epigastric pain in FD patients given CCK-8.16 Cholecystokinin binds to receptors on vagal afferents, and mucosal receptors, in the stomach and small intestine, to potentiate gastric relaxation, stimulate mechanoreceptors sensitive to gastric stretch, and slow gastric emptying.4 Interaction with receptors in satiation centers in the hypothalamus and hindbrain also reduces appetite.8 Cholecystokinin, therefore, acts synergistically with mechanical gastric distension to induce feelings of fullness in response to food.

Hypersensitivity to, or hypersecretion of, other gut hormones which are released in response to food and may play a role in satiation, including glucagon-like peptide-1 (GLP-1) and PYY, has also been postulated to be important in the pathogenesis of dyspeptic symptoms in FD, but evidence is lacking. Like CCK, GLP-1 and PYY are secreted from enteroendocrine cells of the small intestine, especially by fat, and may suppress appetite, slow gastric emptying, and inhibit small intestinal motility.4 Exogenous administration of GLP-1 and PYY decreases energy intake and induces nausea in healthy subjects in a dose-dependent fashion;8 however, the effects of exogenous GLP-1 and PYY have not been evaluated in FD patients. A recent study,13 however, suggests that both fasting and postprandial PYY levels are less, rather than greater, in FD patients than healthy subjects, arguing against a pathogenetic role. Despite this, it remains possible that the relative elevation of plasma PYY concentrations after a high-fat meal compared to a high-carbohydrate meal contributes to the greater severity of dyspeptic symptoms associated with the former.

Gut hormones secreted during fasting which may accelerate gastric emptying, such as motilin and ghrelin, are potential targets for the pathogenesis and management of FD (Table 2). Although fasting motilin levels in FD patients are comparable to those in healthy subjects, the inhibition of proximal gastric accommodation (which is already frequently impaired in FD) by exogenous motilin may be greater in the former.9 Motilin receptor agonists have been used in the management of FD with, and without, gastroparesis, primarily with the rationale that acceleration of gastric emptying may be beneficial. While mitemcinal was reported to accelerate gastric emptying and relieve nausea, fullness and abdominal pain in patients with gastroparesis,17 another motilin agonist, ABT-229, failed to relieve symptoms in patients with FD, possibly as a result of desensitization of motilin receptors and impairment of gastric fundic accommodation,18 and the overall evidence to suggest that motilin receptor agonists will be effective in FD is weak. The role of ghrelin remains uncertain. Functional dyspepsia patients have been reported to have higher,19 lower or unchanged10 fasting ghrelin levels, and normal postprandial ghrelin suppression.13 Acute administration of intravenous ghrelin to anorexic FD patients was reported to stimulate appetite,20 and the ghrelin agonist, TZP-101, has been shown to relieve upper gastrointestinal symptoms and accelerate gastric emptying in diabetic patients with gastroparesis.21 While there is potential for pharmacological doses of ghrelin and/or ghrelin agonists to be used in the management of FD as discussed, there is little evidence that acceleration of gastric emptying per se is beneficial in FD.

Table 2.   Effects of medications that utilize GI hormone signaling on gastric emptying, glycemic control and appetite
MedicationGut hormone actionGastric emptyingGlycemic controlAppetite
  1. GLP-1, glucagon-like peptide-1.

MitemcinalMotilin receptor agonistAccelerates↑ postprandial glucoseNo effect
TZP-101Ghrelin receptor agonistAcceleratesNot known
DexloxiglumideCCK1-receptor antagonistAcceleratesNot knownAttenuates CCK-mediated inhibition of food intake
Exenatide, Exenatide LARGLP-1 analogueSlowsReduces fasting glucose, postprandial glucose excursions and glycated hemoglobin

Impact of gastric emptying on glycemia in diabetes

The rate of gastric emptying is central to postprandial blood glucose homeostasis: the emptying of nutrients from the stomach to the small intestine both influences, and is influenced by, blood glucose concentrations in healthy subjects and individuals with type 1 or type 2 diabetes. The effects of acute changes in the blood glucose concentration on gut motility have been studied extensively since the 1970s. Marked hyperglycemia [blood glucose level ≥15 mmol L−1 (270 mg dL−1)] has a variable effect to slow emptying of solids and nutrient-containing liquids in both type 1 and type 2 diabetic patients.22 Gastric emptying is also affected by variations in blood glucose that are within the normal postprandial range, so that in both healthy subjects and patients with uncomplicated type 1 diabetes, gastric emptying is slower when the blood glucose is ∼8 mmol L−1 (140 mg dL−1) when compared with 4 mmol L−1 (70 mg dL−1).23 In contrast to the effect of acute hyperglycemia, insulin-induced hypoglycemia accelerates gastric emptying substantially,24 which serves as a counter-regulatory mechanism by increasing small intestinal nutrient delivery. The mechanisms mediating the effects of acute hyperglycemia on gastric emptying remain poorly defined, but are likely to involve nitric oxide pathways.25 Acute hyperglycemia has also been shown to affect autonomic function.26

It is not well-recognized that because people in developed affluent countries consume an average of 2000–2500 kcal (and often more!) daily, predominantly as two to three meals which empty from the stomach at 1–4 kcal min−1,2,3 the majority of each day is spent in the ‘postprandial’ or ‘postabsorptive’ states, with only perhaps 3–4 h of true fasting before breakfast. In patients with diabetic gastroparesis, the period of fasting is likely to be even shorter. Moreover, postprandial blood glucose concentrations are a major determinant of diurnal hyperglycemia and overall glycemic control in patients with diabetes, and the contribution of postprandial blood glucose to overall glycemic control increases as glycated hemoglobin decreases below 8.5%.27 Given that glycated hemoglobin is a strong predictor of the microvascular, and possibly macrovascular, complications of diabetes, it is not surprising that there is an increased focus on strategies to minimize postprandial glycemic excursions as part of the management of type 1 and type 2 diabetes.

It has been recognized for some time that the gastric emptying rate accounts for 30–40% of the variance in the initial glycemic response to 75 g oral liquid glucose loads in healthy subjects,28 and in type 2 diabetic patients.29 While gastric emptying has the capacity to affect the total glycemic response to a meal, the magnitude of this effect is less certain. The rate of gastric emptying also influences the initial glycemic response to solid carbohydrate-containing meals.30 Accordingly, in type 2 patients, slowing of gastric emptying by intravenous morphine, and its acceleration by erythromycin, results in a reduction, and an increase, respectively, in both the peak blood glucose and overall glycemic response to a meal,31 and when gastric emptying of a high-fat meal is accelerated by the lipase inhibitor, orlistat, the postprandial glycemic response is greater.32 Similarly, in patients with cystic fibrosis and pancreatic exocrine insufficiency, gastric emptying of meals high in fat and carbohydrate is accelerated, leading to postprandial hyperglycemia.33

Recent studies have established that the relationship between glycemia and small intestinal glucose delivery is non-linear in both healthy subjects34 (Fig. 1) and type 2 diabetic patients.35 In both groups, intraduodenal glucose infusion at a rate of 1 kcal min−1 (i.e., at the lower end of the normal gastric emptying rate) results in a modest increase in blood glucose, whereas the glycemic responses to infusion rates of 2 kcal min−1 and 4 kcal min−1 are substantially greater, but comparable to each other, because of greater plasma insulin response to the 4 kcal min−1 glucose infusion. In health, up to 70% of insulin secretion after an enteral glucose load reflects the so-called ‘incretin effect’,36 which is mediated by the gut hormones, glucagon-like peptide-1 (GLP-1) and glucose-dependent insulinotropic polypeptide (GIP). Glucagon-like peptide-1 and GIP are secreted by specialized enteroendocrine cells in the intestine in response to the presence of ingested nutrients (particularly glucose, and fatty acids derived from lipid digestion). Glucagon-like peptide-1 is found predominantly in the distal small intestine and colon, while GIP is secreted primarily from the proximal small intestine.4 In poorly controlled type 2 diabetes the incretin effect is markedly reduced; this may represent an epiphenomenon, and appears to be primarily because the capacity of GIP to stimulate insulin is impaired by hyperglycemia. The evidence for reduced GLP-1 secretion is mixed.37 When glucose is given intraduodenally at rates within the normal range for gastric emptying in healthy subjects34and type 2 patients,35 the GIP response is load-dependent, with an initial rapid increase followed by a plateau. In contrast, the GLP-1 rise in response to glucose levels of 1 kcal min−1 and 2 kcal min−1 is modest, and substantially greater in response to 4 kcal min−1 (Fig. 2). Glucagon-like peptide-1-induced insulin secretion is likely to account for the relatively small difference in postload glucose excursions at 2 kcal min−1 and 4 kcal min−1. The implications of these observations are that GLP-1 is of particular importance at higher rates of duodenal glucose delivery, and when the rate of gastric emptying is ∼1 kcal min−1 in health and type 2 diabetes there is relatively little rise in blood glucose. Studies employing the specific GLP-1 receptor antagonist, exendin 9–39, indicate that endogenous GLP-1 regulates gastric emptying38and gastroduodenal motility.39 Pharmacological doses of GLP-1 cause a substantial, dose-related, slowing of gastric emptying in health,40 and type 2 diabetes,41 with consequent reductions in postprandial glycemia and insulinemia. Hence, inhibition of gastric emptying by exogenous GLP-1 may outweigh its direct, insulinotropic, effects. In contrast to GLP-1, GIP does not appear to affect gastric emptying.42 In insulin-treated patients the glycemic response to a meal is also influenced by gastric emptying.

Figure 1.

 Blood glucose (A), plasma insulin (B) and plasma glucagon-like peptide-1 (C) concentrations in response to 120-min intraduodenal glucose infusions at rates of 1 (G1), 2 (G2) or 4 (G4) kcal min−1, or saline control (C) in 10 healthy males. (A) *< 0.05 vs control, #< 0.05 vs G1, §< 0.05 vs G2. (B) *< 0.05 vs control, #< 0.05 vs G1, §< 0.05 vs G2. (C) *< 0.05 vs control, #< 0.05 vs G1, §< 0.05 vs G2. (D) *< 0.05 vs control, #< 0.05 vs G1, §< 0.05 vs G2. Data are mean ± SEM. Reproduced with permission from Pilichiewicz et al.34

Figure 2.

 Plasma peptide YY (PYY, A) and cholecystokinin (CCK, B) concentrations during fasting and in response to duodenal nutrient stimulation in feed-tolerant (= 9) and feed-intolerant (= 10) critically ill patients. Both fasting and nutrient-stimulated PYY and CCK concentrations are higher in feed-intolerant patients. *< 0.05, **< 0.01. Data are mean ± SEM. Reproduced with permission from Nguyen et al.53

It has been recognized for many years that in type 1 patients the insulin requirement to maintain euglycemia in the first ∼2 h after a meal is less in patients with gastroparesis than in those with normal gastric emptying,43 and that delayed gastric emptying occurs in 30–50% of patients with long-standing diabetes.1 In type 1 diabetic patients with delayed gastric emptying, it was reported that plasma ghrelin levels are decreased;44 while it has been suggested that this is of pathogenetic significance, this remains uncertain. There is evidence that delayed gastric emptying is not infrequently associated with recurrent, ‘unexplained’, postprandial hypoglycemia in insulin-treated type 1 and 2 diabetic patients, even in the absence of upper gastrointestinal symptoms.45 Accordingly, measurement of gastric emptying in patients with so-called ‘gastric’ hypoglycemia may facilitate the selection and timing of meals and insulin administration.

The above observations have provided a rationale for the recent development of strategies to improve glycemic control in type 1 and type 2 diabetes by modulating gastric emptying. These strategies differ fundamentally between patients who are treated with insulin (type 1 and an increasing number of type 2), and those who are not (type 2). In the former group the focus must be on optimizing the coordination between gastric emptying and insulin delivery i.e., gastric emptying needs to be predictable. In contrast, in patients who are not treated with insulin, slowing of gastric emptying by dietary or pharmacological means should prove beneficial, as long as this is not associated with the induction of upper gastrointestinal symptoms (which has hitherto not been shown to be the case).

While predictable acceleration, or ‘normalization’ of gastric emptying with the use of prokinetic drugs (such as erythromycin, dompepridone and metoclopramide) is logical in insulin-treated patients, there is little evidence to support the efficacy of this approach. Other strategies for insulin-treated patients with imbalance between nutrient delivery and the timing, or dosage of insulin include the use of insulin lispro and insulin aspart, which have a shorter duration of action when compared to regular human insulin.46

The outcome of recent studies employing duodenal glucose infusion suggests that in non-insulin treated patients it would be desirable for the rate of gastric emptying of carbohydrate to be ≤1 kcal min−1 for there to be minimal rise in postprandial glucose levels.33,35,47,48 Some recent studies have employed a ‘preload’ strategy i.e., ingestion of a relatively small amount of nutrient before a meal with the rationale that gastric emptying of the meal will be slowed by stimulating small intestinal hormonal and neural feedback mechanisms, and incretin hormone and insulin secretion promptly stimulated. An olive oil ‘preload’ slows gastric emptying substantially in type 2 diabetic patients and delays the postprandial rise in blood glucose, but only has a modest effect on the peak blood glucose response.47 In contrast, a whey protein ‘preload’ reduces the peak blood glucose and overall glucose response markedly, probably because in addition to slowing gastric emptying and stimulating incretins, amino acids stimulate insulin secretion.48 Pharmacological therapies which potentiate the incretin effect and slow gastric emptying have now entered the mainstream of therapy for diabetes. The GLP-1 receptor agonist, exenatide, improves glycemic control, both acutely and with chronic administration, in type 2 diabetic patients by the combined effects of glucose-dependent insulin secretion, suppression of glucagon release, and, in the case of postprandial glycemia, predominantly by the slowing of gastric emptying49 (Table 2). The latter may be less marked with the once-weekly exenatide preparation (exenatide LAR) which is in development, potentially because of the development of ‘tachyphylaxis’ i.e., the continuous GLP-1 receptor activation may be associated with ‘down-regulation’ and/or changes in vagal pathways.50 The magnitude of the slowing of gastric emptying induced by exenatide in type 2 patients is most marked in those with relatively faster gastric emptying i.e., the slowing and reduction in glycemic excursions are less marked when gastric emptying is relatively slower.49 It is, accordingly, clear that the ‘chicken-and-egg’ relationship between gastric emptying and postprandial glycemia has important therapeutic applications for the management of glycemic control and symptomatic diabetic gastroparesis. The way to improved glycemic control in diabetes may, indeed, be through the stomach.

Critical illness and gut hormones

Enteral feeding is usually the preferred method of nutritional support in the critically ill, but intolerance to nasogastric feeding often occurs as a result of markedly delayed gastric emptying. The slowing of gastric emptying in the critically ill is associated with increased localized pyloric motility, and a reduction in phasic antral motility, reflecting, at least in part, an increase in small intestinal nutrient-mediated inhibitory feedback on gastric emptying.51 These observations have stimulated an increased interest in humoral abnormalities to account for gastroparesis in the critically ill, where prokinetic therapy forms the mainstay of current therapy. It should, of course, be recognized that there are numerous alternative explanations for delayed gastric emptying in this group, including inflammatory cytokines, other metabolic derangements, and septicemia.

In critically ill patients, fasting plasma concentrations of CCK52and PYY53 and those stimulated by duodenal nutrient infusion52,53 are elevated by approximately two-fold, compared with healthy subjects. In those critically ill patients who have feed intolerance, plasma levels of CCK and PYY are substantially higher than in those without intolerance52,53 (Fig. 2). Moreover, the rate of gastric emptying (as measured by a 13C-breath test) in mechanically ventilated intensive care patients is inversely related to fasting and postprandial CCK and PYY concentrations,54 consistent with the concept that CCK and PYY modulate gastric emptying in critical illness. However, the postprandial increases in plasma CCK and PYY are also directly related to the rate of gastric emptying in this group,54 indicating that secretion of these gut hormones is dependent on nutrient delivery into the small intestine, as is to be expected, and accordingly, the overall postprandial rise in CCK and PYY in patients with delayed gastric emptying is comparable to that in those with normal gastric emptying. It should be recognized that these studies have employed breath tests, rather than the ‘gold standard’ of scintigraphy, to measure gastric emptying and there are several aspects of critical illness (e.g. disruption of the gut barrier, respiratory failure) that have the potential to affect the assumptions and calculations entailed in breath testing. Nevertheless, breath tests do appear to be a valid measure of gastric emptying in the critically ill.55 No studies have, to our knowledge, evaluated the effects of CCK-receptor blockade on gastric emptying in the critically ill. The mechanisms underlying the humoral abnormalities in critical illness remain to be clarified; we speculate that it may reflect ‘hypersensitivity’ of enteroendocrine cells of the small intestine, resulting in a greater release of satiating gut hormones in response to a given nutrient load. Inadequate nutrition may contribute to the changes in gut hormone secretion. While anorexia nervosa should not be equated with critical illness, it is of interest that higher CCK and PYY levels are observed in patients with anorexia nervosa, and elderly subjects with malnutrition, while fasting slows gastric emptying in healthy lean and obese subjects.56 Increased PYY secretion within 20 min of a meal is likely to be attributable to neurohormonal factors in the proximal small intestine, rather than direct nutrient (ileal) stimulation, particularly as prolonged intestinal transit is common in the critically ill. Cholecystokinin, secreted in the duodenum and upper jejunum, is a likely candidate, as it stimulates PYY release in healthy humans.8 That higher fasting and postprandial plasma CCK levels correlate strongly with PYY concentrations in critically ill patients,53 also suggests that the rise in PYY is related to the release of CCK. In mice, small intestinal inflammation, mediated by CD4+ T-lymphocytes and the cytokines interleukin-3 and interleukin-4, leads to up-regulation of CCK-secreting cells and an increase in plasma CCK concentrations, and appetite suppression.57 Similar processes associated with systemic inflammation may contribute to elevated CCK and PYY levels in critically ill humans. Peptide YY release from the distal small intestine could also be triggered by direct stimulation of vagal sensory afferents in the proximal small intestine.8 Iatrogenic factors, including mechanical ventilation and the use of inotropic medications, may also contribute to CCK and PYY hypersecretion.

The above observations indicate that altered secretion of gut hormones, in particular the exaggerated inhibitory feedback response of CCK and PYY, is likely to be of pathogenetic significance in the disordered gastrointestinal motility observed in critically ill patients. Increased levels of CCK and PYY may exacerbate impaired energy intake by slowing gastric emptying in patients in whom nutrition is already compromised. Therapeutic strategies which modulate the actions of these hormones, such as pharmacological antagonists of CCK and PYY, may potentially be useful in the management of cachexia and malnutrition associated with critical illness, and deserve further study.

Conclusion

Gut hormones are important in the regulation of gastric emptying. While the pathogenetic significance of abnormalities of secretion and/or action of these hormones in gastroparesis, FD and feed intolerance in critically ill patients is incompletely understood, studies employing nutrient stimulation, specific antagonists and exogenous administration of enterogastric hormones, suggest that they represent targets for an improved understanding and management of these disease states. In addition, the incretin hormones, GLP-1 and GIP, are involved in the mechanisms by which modulation of gastric emptying improves glycemic control in type 2 diabetic patients. The potential therapeutic applications of gut hormones in gastrointestinal disease warrant further evaluation.

Acknowledgments

JK drafted and revised the manuscript; CKR, CF-B, KLJ, and MH all contributed to the critical review and rewriting of the manuscript.

Competing Interests

The authors have no competing interests.

Ancillary